Steel Deoxidation: Part One

Deoxidation is the removal of excess oxygen from molten metal. The procedure involves adding materials with a high affinity for oxygen, the oxides of which are either gaseous or readily form slags. The deoxidation of steel is usually performed by adding Mn, Si and Al, or rarely by adding Cr, V, Ti, Zr and B.

The deoxidation of molten steel shows a paradox. By increasing the concentration of deoxidizer in the melt over some critical value reoxidation of steel takes place. A few examples of the reoxidation of steel by adding the usual deoxidizers (Si and Al) are examined in this article.

Deoxidation is the last stage in steelmaking. In the Basic Oxygen Furnace (BOF) and other similar steelmaking practices the steel bath as the time of tapping contains 400 to 800 ppm activity of oxygen. Deoxidation is carried out during tapping by adding into the tap-ladle appropriate amounts of ferromanganese, ferrosilicon and/or aluminum or other special deoxidizers. If at the end of the blow the carbon content of the steel is below specifications, the metal is also recarburized in the ladle. However, large additions in the ladle are undesirable, because of the adverse effect on the temperature of the metal.

Eight typical conditions of commercial ingots, cast in identical bottle-top molds, in relation to the degree of suppression of gas evolution are shown schematically in Figure 1. The dotted line indicates the height to which the steel originally was poured in each ingot mold. Depending on the carbon content and particularly of the oxygen content, the ingot structures range from that of a fully killed or dead-killed ingot N°1 to that of a violently rimmed ingot N°8. Included in the series are indicated in figure 1 i.e. killed steel N°1, semikilled steel N°2, capped steel N°5, and rimmed steel N°7.

Figure 1: Series of typical ingot structures

Rimmed steel usually is tapped without having made additions of deoxidizers to the steel in the furnace or only small additions to the molten steel in ladle, in order to have sufficient oxygen present to give the desired gas evolution by reacting in the mold with carbon. The exact procedures followed depend upon whether the steel has a carbon content in the higher ranges i.e. %C=0.12-0.15 or in the lower ranges, e.g. %C=max 0.10 .When the metal in the ingot mold begins to solidify, there is a brisk evolution of carbon monoxide, resulting in an outer ingot skin of relatively clean metal low in carbon and other solutes. Such ingots are best suited for the manufacture of steel sheets.

Capped steel practice is a variation of rimmed steel practice. The rimming action is allowed to begin normally, but is then terminated after a minute or more by sealing the mold with a cast-iron cap. In steels with a carbon content greater than 0.15% the capped ingot practice is usually applied to sheet, strip, wire and bars.

Semikilled steel is deoxidized less than killed steel and there is enough oxygen present in the molten steel to react with carbon forming sufficient carbon monoxide to counterbalance the solidification shrinkage. The steel generally has a carbon content within the range %C=0.15-0.30 and finds wide application in structural shapes.

Killed steel is deoxidized to such an extent that there is no gas evolution during solidification. Aluminum is used for deoxidation, together with ferro-alloys of manganese and silicon; in certain cases calcium silicide or other special strong deoxidizers are used. In order to minimize piping, almost all killed steels are cast in hot-topped big-end up molds.

Killed steels are generally used when a homogeneous structure is required in the finished steels. Alloy steels, forging steels and steels for carburizing are of this type, when the essential quality is soundness. In producing certain extra-deep-drawing steels, a low-carbon (%C=max 0.12) steel is killed, usually with a substantial amount of aluminum that is added in the ladle, in the mold or both.

Although the deoxidation of steel by aluminum suppresses the formation of carbon monoxide during solidification, and hence suppresses blow holes, there are many steel processing operations where aluminum killing of steel is undesirable. For example, it is widely recognized that certain alloy steels to be cast as large ingots should not be subject to aluminum killing, because of the piping and of deleterious effects of alumina inclusions on the subsequent processing of ingots for certain applications, e.g. generator-rotor shafts.

It has been recognized from the early days of the continuous-casting operation nearly two decades ago that casting difficulties and poor surface conditions are often experienced with aluminum-killed steels. It is for these reasons that other forms of deoxidation are often preferred in a number of steel-processing operations, e.g. silico-managanese deoxidation and/or vacuum carbon deoxidation.

Deoxidation equilibria

Deoxidation reactions can be described using the deoxidation equilibrium constant. A wide spectrum of deoxidation equilibria pertaining to the most common deoxidants for steel is summarized in Table 1 as a log-log plot of the concentration of oxygen in solution in liquid steel against that of the added elements.

Table 1: Solubility of the products of deoxidation in liquid iron.

• Activities are chosen such that a_Mn ≡ %Mn and a_O ≡ %O when %M→O
• Square brackets [ ] denote component present in molten steel
• Temperature (T) is on the Kelvin scale.

In all cases, the oxygen and the alloying element in solution are in equilibrium with the appropriate gas, liquid or solid oxide phases at 1600°C, e.g. 1 atm CO, pure B₂O₃, pure Al₂O₃ etc. The curves for Mn, Si and C are from compiled data. The curves for Cr, V, B, Ti and Al are based on the recent work done in this laboratory by Fruehan using the oxygen galvanic cell previously described in measuring the equilibrium oxygen potentials.

Deoxidation reactions can be described using the deoxidation equilibrium constant. The reaction when the alloying element (M) is added to the steel can be represented by:

M_xO_y = xM + Yo .....(1)

The deoxidation constant assuming pure M_xO_y forms (i.e. unit activity for M_xO_y) is given by:

K = (h_M)^x(h_O)^y .....(2)

Where h_M and h_O are the Henrian activities defined such that activity of the components is equal to its weight percent at infinite dilution in iron.

H_i = f_i(wt.% i) .....(3)

The activity coefficient f_i can be corrected for alloying elements by use of the interaction parameter e^j_i

(d log f_i/d log wt%j) = e^j_i .....(4)

Table 2 shows the coefficients of interaction for the common elements of carbon and stainless steels at 1600°C.

Table 2: The coefficients of interaction for the common elements of carbon and stainless steels at 1600°C

For most low alloy steels encountered in ladle metallurgy the activity coefficient can be taken as unity and equation 2 reduces to:

K_M = (%M)^x(%O)^y .....(5)

To illustrate how to use these constants consider a steel containing 0.1%Si at 1600°C (2912°F) in equilibrium with SiO₂. The value of K_Si is given by:

K_Si = (%Si)(%O)² .....(6)

K_Si = 2.2 x 10^-5

Therefore:

(%O)² = 2.2 x 10^-4

(%O) ≈ 0.015 or 150 ppm.

It is important to remember that these calculations are for soluble oxygen content; the total oxygen content which includes both the soluble oxygen and the oxygen associated with inclusions could be much higher.

For single element deoxidation, the solubility of oxygen in liquid iron at 1600°C (2912°F) is given as a function of the concentration of the alloying element. In each case, the melt is in equilibrium with the respective pure oxide; e.g. SiO₂, Al₂O₃ etc. It can be clearly seen that aluminum is the strongest of the common deoxiders followed by titanium. Rare earths are about as strong aluminum as deoxidizers and will be discussed later in detail.